pyDiffTools 0.1.56__tar.gz → 0.1.58__tar.gz

{pydifftools-0.1.56/pyDiffTools.egg-info → pydifftools-0.1.58}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: pyDiffTools
-Version: 0.1.56
+Version: 0.1.58
 Summary: Diff tools
 Author: J M Franck
 License: Copyright (c) 2015, jmfranck
@@ -163,6 +163,17 @@ included are (listed in order of fun/utility):
   ``python .github/scripts/update_comment_filter_history.py``; the release
   workflow also refreshes it before building distributions.
 
+  To run Lua filters before pandoc-crossref, declare an ordered list in
+  the Markdown YAML header::
+
+      filters-before-crossref:
+        - filters/panel-references.lua
+
+  Paths are resolved relative to the Markdown file. Each declared filter
+  runs once, in listed order, and is excluded from the adjacent Lua filters
+  that normally run afterward. Missing files produce a build error.
+  Documents without this setting retain the existing filter order.
+
   Declare a single local ``.bib`` file with ``bibliography: references.bib``
   in the Markdown YAML header; relative paths are resolved beside the
   Markdown file. Without that declaration, ``cpb`` uses a single adjacent

{pydifftools-0.1.56 → pydifftools-0.1.58}/README.rst

@@ -108,6 +108,17 @@ included are (listed in order of fun/utility):
   ``python .github/scripts/update_comment_filter_history.py``; the release
   workflow also refreshes it before building distributions.
 
+  To run Lua filters before pandoc-crossref, declare an ordered list in
+  the Markdown YAML header::
+
+      filters-before-crossref:
+        - filters/panel-references.lua
+
+  Paths are resolved relative to the Markdown file. Each declared filter
+  runs once, in listed order, and is excluded from the adjacent Lua filters
+  that normally run afterward. Missing files produce a build error.
+  Documents without this setting retain the existing filter order.
+
   Declare a single local ``.bib`` file with ``bibliography: references.bib``
   in the Markdown YAML header; relative paths are resolved beside the
   Markdown file. Without that declaration, ``cpb`` uses a single adjacent

{pydifftools-0.1.56 → pydifftools-0.1.58/pyDiffTools.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: pyDiffTools
-Version: 0.1.56
+Version: 0.1.58
 Summary: Diff tools
 Author: J M Franck
 License: Copyright (c) 2015, jmfranck
@@ -163,6 +163,17 @@ included are (listed in order of fun/utility):
   ``python .github/scripts/update_comment_filter_history.py``; the release
   workflow also refreshes it before building distributions.
 
+  To run Lua filters before pandoc-crossref, declare an ordered list in
+  the Markdown YAML header::
+
+      filters-before-crossref:
+        - filters/panel-references.lua
+
+  Paths are resolved relative to the Markdown file. Each declared filter
+  runs once, in listed order, and is excluded from the adjacent Lua filters
+  that normally run afterward. Missing files produce a build error.
+  Documents without this setting retain the existing filter order.
+
   Declare a single local ``.bib`` file with ``bibliography: references.bib``
   in the Markdown YAML header; relative paths are resolved beside the
   Markdown file. Without that declaration, ``cpb`` uses a single adjacent

{pydifftools-0.1.56 → pydifftools-0.1.58}/pydifftools/browser_lifecycle.py

@@ -4,8 +4,7 @@ import shutil
 import subprocess
 import sys
 
-# callbacks that run before any dialog needs the user; cpb registers one
-# that closes its live preview, so issues are dealt with before it reopens
+# Optional callbacks for callers that need to prepare for a dialog.
 dialog_callbacks = []
 
 
@@ -97,7 +96,7 @@ def start_browser():
 
 
 def prepare_for_dialog():
-    """Let watch commands close their preview before a dialog appears."""
+    """Notify registered callers before a dialog appears."""
     for callback in list(dialog_callbacks):
         callback()
 
@@ -108,6 +107,10 @@ def browser_window_is_alive(browser):
     # the same shutdown behavior when a user closes the browser window.
     # Do not probe with execute_script here: page navigations can briefly
     # interrupt script execution even while the window is still open.
+    from selenium.common.exceptions import (
+        InvalidSessionIdException, NoSuchWindowException, WebDriverException,
+    )
+
     if browser is None:
         return False
     try:
@@ -115,10 +118,14 @@ def browser_window_is_alive(browser):
         if not handles:
             return False
         return True
-    except Exception:
+    except (InvalidSessionIdException, NoSuchWindowException):
         return False
+    except WebDriverException:
+        # A transient command failure is not evidence of a closed window.
+        return True
 
 
+# also used by: continuous.py, notebook/fast_build.py, flowchart/watch_graph.py
 def close_browser_window(browser):
     # Close a Selenium browser session and ignore errors from already-closed
     # windows so cleanup paths stay simple.
@@ -137,20 +144,93 @@ def forward_search_in_browser(browser, search_text):
     if browser is None or not search_text:
         return False
     found = browser.execute_script(
-        """
-        var searchText = arguments[0];
-        if (!window.find) {
-            return false;
+        r"""
+        const searchText = arguments[0];
+        let didFind = window.find && window.find(searchText);
+        // Match across inline elements, entities, and rendered references.
+        if (!didFind) {
+            const escape = text => text.replace(
+                /[.*+?^${}()|[\]\\]/g, '\\$&'
+            ).replace(/\s+/g, '\\s+');
+            const reference =
+                /\[?@([a-z]{2,4}):(\w(?:[\w:./-]*[\w/-])?)\]?/gi;
+            let pattern = '', cursor = 0, match;
+            let valid = true;
+            while ((match = reference.exec(searchText))) {
+                pattern += escape(searchText.slice(cursor, match.index));
+                const id = match[1].toLowerCase() + ':' + match[2];
+                const link = Array.from(
+                    document.querySelectorAll('a[href]')
+                ).find(a => a.getAttribute('href') === '#' + id);
+                const target = document.getElementById(id);
+                if (!link && !target) {
+                    valid = false;
+                    break;
+                }
+                // Unlinked crossrefs still render a numeric label. For a
+                // reference alone, select its target rather than guessing
+                // which occurrence of that number the user intended.
+                if (!link && searchText.trim() === match[0]) {
+                    const range = document.createRange();
+                    range.selectNodeContents(target);
+                    const selection = window.getSelection();
+                    selection.removeAllRanges();
+                    selection.addRange(range);
+                    didFind = true;
+                    valid = false;
+                    break;
+                }
+                pattern += '(?:(?:eq(?:uation)?|fig(?:ure)?|tab(?:le)?|' +
+                    'sec(?:tion)?|ch(?:apter)?|lst|listing)s?\\.?\\s*)?' +
+                    (link ? escape(link.textContent.trim()) :
+                        '\\(?\\d+(?:[.–-]\\d+)*[a-z]?\\)?');
+                cursor = reference.lastIndex;
+            }
+            pattern += escape(searchText.slice(cursor));
+            if (valid) {
+                const walker = document.createTreeWalker(
+                    document.body, NodeFilter.SHOW_TEXT
+                );
+                const nodes = [], starts = [];
+                let text = '', node;
+                while ((node = walker.nextNode())) {
+                    if (node.parentElement.closest(
+                        'script, style, noscript, [hidden]'
+                    ) || !node.parentElement.getClientRects().length) {
+                        continue;
+                    }
+                    starts.push(text.length);
+                    nodes.push(node);
+                    text += node.textContent;
+                }
+                const result = new RegExp(pattern, 'i').exec(text);
+                if (result && result[0].length) {
+                    const start = result.index;
+                    const end = start + result[0].length;
+                    const first = starts.findIndex((offset, i) =>
+                        offset + nodes[i].length > start
+                    );
+                    const last = starts.findIndex((offset, i) =>
+                        offset + nodes[i].length >= end
+                    );
+                    const range = document.createRange();
+                    range.setStart(nodes[first], start - starts[first]);
+                    range.setEnd(nodes[last], end - starts[last]);
+                    const selection = window.getSelection();
+                    selection.removeAllRanges();
+                    selection.addRange(range);
+                    didFind = true;
+                }
+            }
         }
-        var didFind = window.find(searchText);
         if (didFind && window.getSelection) {
-            var selection = window.getSelection();
+            const selection = window.getSelection();
             if (selection.rangeCount > 0) {
-                var rect = selection.getRangeAt(0).getBoundingClientRect();
+                const rect = selection.getRangeAt(0).getBoundingClientRect();
                 window.scrollBy(0, rect.top - window.innerHeight / 3);
             }
         }
-        return didFind;
+        return Boolean(didFind);
         """,
         search_text,
     )

{pydifftools-0.1.56 → pydifftools-0.1.58}/pydifftools/command_line.py

@@ -1,6 +1,7 @@
 # PYTHON_ARGCOMPLETE_OK
 
 import argparse
+import html
 import sys
 import os
 import gzip
@@ -554,26 +555,14 @@ def fs(arguments):
 )
 def mfs(text):
     # Normalize markdown-specific markup so search phrases match rendered text.
-    search_text = text
-    marker_match = re.search(r"(?i)@[a-z]{2,4}:", search_text)
-    if marker_match:
-        search_text = search_text[: marker_match.start()]
-    search_text = re.sub(r"\[@[^\]]+\]", " ", search_text)
-    # Strip paired markup delimiters (bold/italic, strikethrough,
-    # superscript/subscript). These delimiters hug the text they wrap with no
-    # space in the rendered output (e.g. "x^2^" renders as "x2"), so they are
-    # dropped rather than replaced with a space. Plain str.replace is used
-    # (not regex) so this is linear time regardless of how many delimiter
-    # characters appear.
-    search_text = (
-        search_text.replace("**", "")
-        .replace("__", "")
-        .replace("~~", "")
-        .replace("*", "")
-        .replace("_", "")
-        .replace("~", "")
-        .replace("^", "")
-    )
+    search_text = html.unescape(text)
+    search_text = re.sub(r"\[@(?![a-z]{2,4}:)[^\]]+\]", " ", search_text,
+                         flags=re.IGNORECASE)
+    # Keep reference IDs intact while removing surrounding Markdown markup.
+    parts = re.split(r"(?i)(@[a-z]{2,4}:\w(?:[\w:./-]*[\w/-])?)", search_text)
+    for index in range(0, len(parts), 2):
+        parts[index] = re.sub(r"[*_~^]", "", parts[index])
+    search_text = "".join(parts)
     search_text = re.sub(r"\s+", " ", search_text).strip()
     if not search_text:
         search_text = text.strip()
@@ -620,7 +609,8 @@ def mfs(text):
             if not filename.endswith(".md"):
                 continue
             with open(filename, encoding="utf-8") as fp:
-                if search_text in fp.read():
+                source_text = fp.read()
+                if text.strip() in source_text or search_text in source_text:
                     matching_files.append(filename)
         if len(matching_files) == 0:
             raise RuntimeError(

{pydifftools-0.1.56 → pydifftools-0.1.58}/pydifftools/continuous.py

@@ -11,6 +11,7 @@ import re
 import shutil
 import threading
 import queue
+import signal
 import traceback
 import tempfile
 from pathlib import Path
@@ -27,7 +28,6 @@ from .comment_migration import prepare_comment_source
 from .browser_lifecycle import (
     browser_window_is_alive,
     close_browser_window,
-    dialog_callbacks,
     forward_search_in_browser,
     prepare_for_dialog,
     start_browser,
@@ -826,6 +826,7 @@ def run_pandoc(
         # }}}
     # {{{ select the declared bibliography before adjacent-file discovery
     bibliography_declared = False
+    metadata = None
     bibliography_paths = []
     front_matter = re.match(
         r"\A\ufeff?---[^\S\n]*\n(.*?)\n(?:---|\.\.\.)[^\S\n]*(?:\n|$)",
@@ -851,6 +852,27 @@ def run_pandoc(
                     for value in declared
                 ]
     # }}}
+    # {{{ resolve explicitly ordered filters before cross-reference processing
+    filters_before_crossref = []
+    if isinstance(metadata, dict) and "filters-before-crossref" in metadata:
+        declared_filters = metadata["filters-before-crossref"]
+        if not isinstance(declared_filters, list) or not all(
+            isinstance(value, str) and value for value in declared_filters
+        ):
+            raise ValueError(
+                "cpb: filters-before-crossref must be a list of file paths "
+                f"in {filename}"
+            )
+        for value in declared_filters:
+            filter_path = Path(source_dir, value).resolve()
+            if not filter_path.is_file():
+                raise FileNotFoundError(
+                    "cpb: filters-before-crossref file not found: "
+                    f"{value!r} (resolved to {filter_path})"
+                )
+            if str(filter_path) not in filters_before_crossref:
+                filters_before_crossref.append(str(filter_path))
+    # }}}
     localfiles = {}
     for k in ["csl", "bib"]:
         if k == "bib" and bibliography_declared:
@@ -886,7 +908,12 @@ def run_pandoc(
         "author-info-blocks.lua": 1,
     }
     localfiles["lua"] = sorted(
-        [f for f in os.listdir(source_dir) if f.endswith(".lua")],
+        [
+            f for f in os.listdir(source_dir)
+            if f.endswith(".lua")
+            and str(Path(source_dir, f).resolve())
+            not in filters_before_crossref
+        ],
         key=lambda name: (lua_priority.get(name, 2), name),
     )
     # Include any javascript files next to the markdown source by injecting
@@ -908,6 +935,10 @@ def run_pandoc(
         html_file,
         filename,
     ]
+    before_crossref_args = []
+    for filter_path in filters_before_crossref:
+        before_crossref_args.extend(["--lua-filter", filter_path])
+    command[1:1] = before_crossref_args
     if not bibliography_declared and localfiles["bib"]:
         bibliography_paths = [localfiles["bib"]]
     for bibliography_path in bibliography_paths:
@@ -1200,11 +1231,16 @@ def cpb(
         daemon=True,
     )
     chrome = None
-    scroll_position = 0
     observer = None
     observer_started = False
     socket_thread_started = False
     source_jump_server = None
+    previous_sigterm = None
+    if threading.current_thread() is threading.main_thread():
+        def stop_cpb(_signum, _frame):
+            raise KeyboardInterrupt
+
+        previous_sigterm = signal.signal(signal.SIGTERM, stop_cpb)
 
     try:
         # Bind and serve before any build or browser work. This makes the fixed
@@ -1290,29 +1326,18 @@ def cpb(
         # and never be reported as a change.
         chrome.get("file://" + os.path.abspath(html_file))
 
-        def close_for_dialog():
-            # a rebuild needs the user, so close the preview while the
-            # issues are dealt with; it reopens once the rebuild is done
-            nonlocal chrome, scroll_position
-            if chrome is not None:
-                # sessionStorage does not survive quitting the browser.
-                scroll_position = chrome.execute_script(
-                    "return window.scrollY;"
-                )
-                close_browser_window(chrome)
-                chrome = None
-
-        dialog_callbacks.append(close_for_dialog)
         rebuild_pending = False
         stable_signature = None
         stable_since = None
+        listener_warning_shown = False
         while True:
-            if not socket_thread.is_alive():
-                raise RuntimeError(
-                    "The cpb forward-search listener stopped unexpectedly; "
-                    "closing the preview instead of leaving an "
-                    "undiscoverable session."
+            if not socket_thread.is_alive() and not listener_warning_shown:
+                print(
+                    "pydifft cpb: forward-search listener stopped; "
+                    "forward search is unavailable. The preview remains open.",
+                    file=sys.stderr,
                 )
+                listener_warning_shown = True
             # Browser closure is terminal. Source-file and build failures do
             # not affect browser ownership or observer lifetime.
             if not browser_window_is_alive(chrome):
@@ -1321,7 +1346,11 @@ def cpb(
             for queued_search in drain_forward_search_queue(search_queue):
                 search_text = queued_search.strip()
                 if search_text:
-                    forward_search_in_browser(chrome, search_text)
+                    try:
+                        forward_search_in_browser(chrome, search_text)
+                    except WebDriverException as exc:
+                        print(f"pydifft cpb: forward search failed: {exc}",
+                              file=sys.stderr)
 
             saw_source_event = False
             while not change_queue.empty():
@@ -1419,30 +1448,16 @@ def cpb(
                                 traceback.print_exc()
                         else:
                             handled_signature = attempted_signature
-                            if chrome is not None:
-                                if not browser_window_is_alive(chrome):
-                                    break
-                                try:
-                                    chrome.refresh()
-                                except WebDriverException:
-                                    print(
-                                        "pydifft cpb: Browser is no longer "
-                                        "available; stopping without "
-                                        "reopening it.",
-                                        file=sys.stderr,
-                                    )
-                                    break
-                        if chrome is None:
-                            # {{{ reopen the preview a dialog closed
-                            chrome = start_browser()
-                            chrome.get(
-                                "file://" + os.path.abspath(html_file)
-                            )
-                            chrome.execute_script(
-                                "window.scrollTo(0, arguments[0]);",
-                                scroll_position,
-                            )
-                            # }}}
+                            if not browser_window_is_alive(chrome):
+                                break
+                            try:
+                                chrome.refresh()
+                            except WebDriverException as exc:
+                                print(
+                                    "pydifft cpb: preview refresh failed: "
+                                    f"{exc}",
+                                    file=sys.stderr,
+                                )
             # }}}
             time.sleep(MAIN_LOOP_INTERVAL_SECONDS)
     except KeyboardInterrupt:
@@ -1458,7 +1473,8 @@ def cpb(
         if socket_thread_started:
             socket_thread.join()
         close_browser_window(chrome)
-        dialog_callbacks.clear()
+        if previous_sigterm is not None:
+            signal.signal(signal.SIGTERM, previous_sigterm)
 
 
 if __name__ == "__main__":

{pydifftools-0.1.56 → pydifftools-0.1.58}/pydifftools/wrap_sentences.py

@@ -703,7 +703,14 @@ def wrap_prose(
     """Wrap prose with sentence and punctuation rules for both commands."""
     output = []
     # Keep paragraph separators and the boundary newlines next to exclusions.
-    for paragraph in PARAGRAPH_SPLIT.split(text):
+    paragraphs = PARAGRAPH_SPLIT.split(text)
+    if filetype == "markdown":
+        # Hard line breaks are source boundaries, not wrappable whitespace.
+        paragraphs = [
+            part for paragraph in paragraphs
+            for part in re.split(r"( {2,}\r?\n)", paragraph)
+        ]
+    for paragraph in paragraphs:
         if not paragraph.strip():
             output.append(paragraph)
             continue
@@ -1001,7 +1008,9 @@ def check_prose(
             if k < len(separators):
                 sep = separators[k]
                 offset += len(sep)
-                if sep == " " and (
+                if sep == " " and not re.match(
+                    r"[ \t]*\r?\n", stripped[offset:]
+                ) and (
                     k + 1 >= len(sentences) or sentences[k + 1].strip()
                 ):
                     issue_line = para_line + stripped.count(

{pydifftools-0.1.56 → pydifftools-0.1.58}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "pyDiffTools"
-version = "0.1.56"
+version = "0.1.58"
 authors = [{ name = "J M Franck" }]
 description = "Diff tools"
 readme = "README.rst"

{pydifftools-0.1.56 → pydifftools-0.1.58}/tests/test_browser_lifecycle.py

@@ -271,3 +271,85 @@ def test_start_browser_does_not_hide_programming_errors(monkeypatch):
     monkeypatch.setattr(webdriver, "Firefox", firefox)
     with pytest.raises(TypeError, match="unexpected argument"):
         browser_lifecycle.start_browser()
+
+
+def test_forward_search_matches_entities_and_rendered_crossrefs(
+    monkeypatch, tmp_path
+):
+    from selenium.webdriver.chrome.service import Service
+    from pydifftools import command_line
+
+    options = webdriver.ChromeOptions()
+    options.add_argument("--headless=new")
+    browser = webdriver.Chrome(
+        service=Service("/usr/bin/chromedriver"), options=options
+    )
+    monkeypatch.setattr(browser_lifecycle.shutil, "which", lambda _name: None)
+    searches = []
+    monkeypatch.setattr(
+        command_line, "send_forward_search",
+        lambda _address, text: searches.append(text),
+    )
+    try:
+        preview = tmp_path / "preview.html"
+        preview.write_text(
+            "<p>It melts near 238&nbsp;K &amp; stays solid.</p>"
+            '<p>See eq.&nbsp;<a href="#eq:my_equation">(1)</a> for '
+            "<strong>the result</strong>.</p>"
+            '<p><a href="#fig:overview">Figure 2</a> shows the sample.</p>'
+            '<p>See <a href="#sec:methods">section 3</a>.</p>'
+            '<h2 id="sec:unlinked">4 Unlinked section</h2>'
+            "<p>As described in sec.&nbsp;4, it works.</p>"
+        )
+        browser.get(preview.as_uri())
+        examples = [
+            ("It melts near 238&nbsp;K &amp; stays solid.",
+             "It melts near 238\xa0K & stays solid."),
+            ("238&#160;K", "238\xa0K"),
+            ("238 K", "238\xa0K"),
+            ("See @eq:my_equation for **the result**.",
+             "See eq.\xa0(1) for the result."),
+            ("@eq:my_equation for the result.",
+             "eq.\xa0(1) for the result."),
+            ("See [@eq:my_equation]", "See eq.\xa0(1)"),
+            ("@fig:overview shows the sample.",
+             "Figure 2 shows the sample."),
+            ("See @sec:methods.", "See section 3."),
+            ("As described in @sec:unlinked, it works.",
+             "As described in sec.\xa04, it works."),
+            ("@sec:unlinked", "4 Unlinked section"),
+        ]
+        for source, rendered in examples:
+            browser.execute_script("window.getSelection().removeAllRanges()")
+            command_line.mfs(source)
+            assert browser_lifecycle.forward_search_in_browser(
+                browser, searches[-1]
+            ), source
+            assert browser.execute_script(
+                "return window.getSelection().toString()"
+            ) == rendered
+        assert not browser_lifecycle.forward_search_in_browser(
+            browser, "See @eq:missing for the result."
+        )
+    finally:
+        browser.quit()
+
+
+@pytest.mark.parametrize("closed", [False, True])
+def test_browser_probe_distinguishes_transient_error_from_closed_session(
+    closed,
+):
+    class InterruptedBrowser:
+        @property
+        def window_handles(self):
+            if closed:
+                from selenium.common.exceptions import (
+                    InvalidSessionIdException,
+                )
+
+                raise InvalidSessionIdException("window closed")
+            raise WebDriverException("page navigation interrupted command")
+
+    assert browser_lifecycle.browser_window_is_alive(
+        InterruptedBrowser()
+    ) is not closed